[net-next-2.6.git] / net / ipv6 / anycast.c

/*
 *	Anycast support for IPv6
 *	Linux INET6 implementation
 *
 *	Authors:
 *	David L Stevens (dlstevens@us.ibm.com)
 *
 *	based heavily on net/ipv6/mcast.c
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/capability.h>
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/random.h>
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/route.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>

#include <net/sock.h>
#include <net/snmp.h>

#include <net/ipv6.h>
#include <net/protocol.h>
#include <net/if_inet6.h>
#include <net/ndisc.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>

#include <net/checksum.h>

static int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr);

/* Big ac list lock for all the sockets */
static DEFINE_RWLOCK(ipv6_sk_ac_lock);

static int
ip6_onlink(struct in6_addr *addr, struct net_device *dev)
{
	struct inet6_dev	*idev;
	struct inet6_ifaddr	*ifa;
	int	onlink;

	onlink = 0;
	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev) {
		read_lock_bh(&idev->lock);
		for (ifa=idev->addr_list; ifa; ifa=ifa->if_next) {
			onlink = ipv6_prefix_equal(addr, &ifa->addr,
						   ifa->prefix_len);
			if (onlink)
				break;
		}
		read_unlock_bh(&idev->lock);
	}
	rcu_read_unlock();
	return onlink;
}

/*
 *	socket join an anycast group
 */

int ipv6_sock_ac_join(struct sock *sk, int ifindex, struct in6_addr *addr)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev = NULL;
	struct inet6_dev *idev;
	struct ipv6_ac_socklist *pac;
	int	ishost = !ipv6_devconf.forwarding;
	int	err = 0;

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;
	if (ipv6_addr_is_multicast(addr))
		return -EINVAL;
	if (ipv6_chk_addr(addr, NULL, 0))
		return -EINVAL;

	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
	if (pac == NULL)
		return -ENOMEM;
	pac->acl_next = NULL;
	ipv6_addr_copy(&pac->acl_addr, addr);

	if (ifindex == 0) {
		struct rt6_info *rt;

		rt = rt6_lookup(addr, NULL, 0, 0);
		if (rt) {
			dev = rt->rt6i_dev;
			dev_hold(dev);
			dst_release(&rt->u.dst);
		} else if (ishost) {
			err = -EADDRNOTAVAIL;
			goto out_free_pac;
		} else {
			/* router, no matching interface: just pick one */

			dev = dev_get_by_flags(IFF_UP, IFF_UP|IFF_LOOPBACK);
		}
	} else
		dev = dev_get_by_index(ifindex);

	if (dev == NULL) {
		err = -ENODEV;
		goto out_free_pac;
	}

	idev = in6_dev_get(dev);
	if (!idev) {
		if (ifindex)
			err = -ENODEV;
		else
			err = -EADDRNOTAVAIL;
		goto out_dev_put;
	}
	/* reset ishost, now that we have a specific device */
	ishost = !idev->cnf.forwarding;
	in6_dev_put(idev);

	pac->acl_ifindex = dev->ifindex;

	/* XXX
	 * For hosts, allow link-local or matching prefix anycasts.
	 * This obviates the need for propagating anycast routes while
	 * still allowing some non-router anycast participation.
	 */
	if (!ip6_onlink(addr, dev)) {
		if (ishost)
			err = -EADDRNOTAVAIL;
		if (err)
			goto out_dev_put;
	}

	err = ipv6_dev_ac_inc(dev, addr);
	if (err)
		goto out_dev_put;

	write_lock_bh(&ipv6_sk_ac_lock);
	pac->acl_next = np->ipv6_ac_list;
	np->ipv6_ac_list = pac;
	write_unlock_bh(&ipv6_sk_ac_lock);

	dev_put(dev);

	return 0;

out_dev_put:
	dev_put(dev);
out_free_pac:
	sock_kfree_s(sk, pac, sizeof(*pac));
	return err;
}

/*
 *	socket leave an anycast group
 */
int ipv6_sock_ac_drop(struct sock *sk, int ifindex, struct in6_addr *addr)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev;
	struct ipv6_ac_socklist *pac, *prev_pac;

	write_lock_bh(&ipv6_sk_ac_lock);
	prev_pac = NULL;
	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
		if ((ifindex == 0 || pac->acl_ifindex == ifindex) &&
		     ipv6_addr_equal(&pac->acl_addr, addr))
			break;
		prev_pac = pac;
	}
	if (!pac) {
		write_unlock_bh(&ipv6_sk_ac_lock);
		return -ENOENT;
	}
	if (prev_pac)
		prev_pac->acl_next = pac->acl_next;
	else
		np->ipv6_ac_list = pac->acl_next;

	write_unlock_bh(&ipv6_sk_ac_lock);

	dev = dev_get_by_index(pac->acl_ifindex);
	if (dev) {
		ipv6_dev_ac_dec(dev, &pac->acl_addr);
		dev_put(dev);
	}
	sock_kfree_s(sk, pac, sizeof(*pac));
	return 0;
}

void ipv6_sock_ac_close(struct sock *sk)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev = NULL;
	struct ipv6_ac_socklist *pac;
	int	prev_index;

	write_lock_bh(&ipv6_sk_ac_lock);
	pac = np->ipv6_ac_list;
	np->ipv6_ac_list = NULL;
	write_unlock_bh(&ipv6_sk_ac_lock);

	prev_index = 0;
	while (pac) {
		struct ipv6_ac_socklist *next = pac->acl_next;

		if (pac->acl_ifindex != prev_index) {
			if (dev)
				dev_put(dev);
			dev = dev_get_by_index(pac->acl_ifindex);
			prev_index = pac->acl_ifindex;
		}
		if (dev)
			ipv6_dev_ac_dec(dev, &pac->acl_addr);
		sock_kfree_s(sk, pac, sizeof(*pac));
		pac = next;
	}
	if (dev)
		dev_put(dev);
}

#if 0
/* The function is not used, which is funny. Apparently, author
 * supposed to use it to filter out datagrams inside udp/raw but forgot.
 *
 * It is OK, anycasts are not special comparing to delivery to unicasts.
 */

int inet6_ac_check(struct sock *sk, struct in6_addr *addr, int ifindex)
{
	struct ipv6_ac_socklist *pac;
	struct ipv6_pinfo *np = inet6_sk(sk);
	int	found;

	found = 0;
	read_lock(&ipv6_sk_ac_lock);
	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
		if (ifindex && pac->acl_ifindex != ifindex)
			continue;
		found = ipv6_addr_equal(&pac->acl_addr, addr);
		if (found)
			break;
	}
	read_unlock(&ipv6_sk_ac_lock);

	return found;
}

#endif

static void aca_put(struct ifacaddr6 *ac)
{
	if (atomic_dec_and_test(&ac->aca_refcnt)) {
		in6_dev_put(ac->aca_idev);
		dst_release(&ac->aca_rt->u.dst);
		kfree(ac);
	}
}

/*
 *	device anycast group inc (add if not found)
 */
int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
{
	struct ifacaddr6 *aca;
	struct inet6_dev *idev;
	struct rt6_info *rt;
	int err;

	idev = in6_dev_get(dev);

	if (idev == NULL)
		return -EINVAL;

	write_lock_bh(&idev->lock);
	if (idev->dead) {
		err = -ENODEV;
		goto out;
	}

	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
		if (ipv6_addr_equal(&aca->aca_addr, addr)) {
			aca->aca_users++;
			err = 0;
			goto out;
		}
	}

	/*
	 *	not found: create a new one.
	 */

	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);

	if (aca == NULL) {
		err = -ENOMEM;
		goto out;
	}

	rt = addrconf_dst_alloc(idev, addr, 1);
	if (IS_ERR(rt)) {
		kfree(aca);
		err = PTR_ERR(rt);
		goto out;
	}

	ipv6_addr_copy(&aca->aca_addr, addr);
	aca->aca_idev = idev;
	aca->aca_rt = rt;
	aca->aca_users = 1;
	/* aca_tstamp should be updated upon changes */
	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	atomic_set(&aca->aca_refcnt, 2);
	spin_lock_init(&aca->aca_lock);

	aca->aca_next = idev->ac_list;
	idev->ac_list = aca;
	write_unlock_bh(&idev->lock);

	dst_hold(&rt->u.dst);
	if (ip6_ins_rt(rt))
		dst_release(&rt->u.dst);

	addrconf_join_solict(dev, &aca->aca_addr);

	aca_put(aca);
	return 0;
out:
	write_unlock_bh(&idev->lock);
	in6_dev_put(idev);
	return err;
}

/*
 *	device anycast group decrement
 */
int __ipv6_dev_ac_dec(struct inet6_dev *idev, struct in6_addr *addr)
{
	struct ifacaddr6 *aca, *prev_aca;

	write_lock_bh(&idev->lock);
	prev_aca = NULL;
	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
		if (ipv6_addr_equal(&aca->aca_addr, addr))
			break;
		prev_aca = aca;
	}
	if (!aca) {
		write_unlock_bh(&idev->lock);
		return -ENOENT;
	}
	if (--aca->aca_users > 0) {
		write_unlock_bh(&idev->lock);
		return 0;
	}
	if (prev_aca)
		prev_aca->aca_next = aca->aca_next;
	else
		idev->ac_list = aca->aca_next;
	write_unlock_bh(&idev->lock);
	addrconf_leave_solict(idev, &aca->aca_addr);

	dst_hold(&aca->aca_rt->u.dst);
	if (ip6_del_rt(aca->aca_rt))
		dst_free(&aca->aca_rt->u.dst);
	else
		dst_release(&aca->aca_rt->u.dst);

	aca_put(aca);
	return 0;
}

static int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr)
{
	int ret;
	struct inet6_dev *idev = in6_dev_get(dev);
	if (idev == NULL)
		return -ENODEV;
	ret = __ipv6_dev_ac_dec(idev, addr);
	in6_dev_put(idev);
	return ret;
}

/*
 *	check if the interface has this anycast address
 */
static int ipv6_chk_acast_dev(struct net_device *dev, struct in6_addr *addr)
{
	struct inet6_dev *idev;
	struct ifacaddr6 *aca;

	idev = in6_dev_get(dev);
	if (idev) {
		read_lock_bh(&idev->lock);
		for (aca = idev->ac_list; aca; aca = aca->aca_next)
			if (ipv6_addr_equal(&aca->aca_addr, addr))
				break;
		read_unlock_bh(&idev->lock);
		in6_dev_put(idev);
		return aca != 0;
	}
	return 0;
}

/*
 *	check if given interface (or any, if dev==0) has this anycast address
 */
int ipv6_chk_acast_addr(struct net_device *dev, struct in6_addr *addr)
{
	if (dev)
		return ipv6_chk_acast_dev(dev, addr);
	read_lock(&dev_base_lock);
	for (dev=dev_base; dev; dev=dev->next)
		if (ipv6_chk_acast_dev(dev, addr))
			break;
	read_unlock(&dev_base_lock);
	return dev != 0;
}


#ifdef CONFIG_PROC_FS
struct ac6_iter_state {
	struct net_device *dev;
	struct inet6_dev *idev;
};

#define ac6_seq_private(seq)	((struct ac6_iter_state *)(seq)->private)

static inline struct ifacaddr6 *ac6_get_first(struct seq_file *seq)
{
	struct ifacaddr6 *im = NULL;
	struct ac6_iter_state *state = ac6_seq_private(seq);

	for (state->dev = dev_base, state->idev = NULL;
	     state->dev;
	     state->dev = state->dev->next) {
		struct inet6_dev *idev;
		idev = in6_dev_get(state->dev);
		if (!idev)
			continue;
		read_lock_bh(&idev->lock);
		im = idev->ac_list;
		if (im) {
			state->idev = idev;
			break;
		}
		read_unlock_bh(&idev->lock);
	}
	return im;
}

static struct ifacaddr6 *ac6_get_next(struct seq_file *seq, struct ifacaddr6 *im)
{
	struct ac6_iter_state *state = ac6_seq_private(seq);

	im = im->aca_next;
	while (!im) {
		if (likely(state->idev != NULL)) {
			read_unlock_bh(&state->idev->lock);
			in6_dev_put(state->idev);
		}
		state->dev = state->dev->next;
		if (!state->dev) {
			state->idev = NULL;
			break;
		}
		state->idev = in6_dev_get(state->dev);
		if (!state->idev)
			continue;
		read_lock_bh(&state->idev->lock);
		im = state->idev->ac_list;
	}
	return im;
}

static struct ifacaddr6 *ac6_get_idx(struct seq_file *seq, loff_t pos)
{
	struct ifacaddr6 *im = ac6_get_first(seq);
	if (im)
		while (pos && (im = ac6_get_next(seq, im)) != NULL)
			--pos;
	return pos ? NULL : im;
}

static void *ac6_seq_start(struct seq_file *seq, loff_t *pos)
{
	read_lock(&dev_base_lock);
	return ac6_get_idx(seq, *pos);
}

static void *ac6_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct ifacaddr6 *im;
	im = ac6_get_next(seq, v);
	++*pos;
	return im;
}

static void ac6_seq_stop(struct seq_file *seq, void *v)
{
	struct ac6_iter_state *state = ac6_seq_private(seq);
	if (likely(state->idev != NULL)) {
		read_unlock_bh(&state->idev->lock);
		in6_dev_put(state->idev);
	}
	read_unlock(&dev_base_lock);
}

static int ac6_seq_show(struct seq_file *seq, void *v)
{
	struct ifacaddr6 *im = (struct ifacaddr6 *)v;
	struct ac6_iter_state *state = ac6_seq_private(seq);

	seq_printf(seq,
		   "%-4d %-15s " NIP6_SEQFMT " %5d\n",
		   state->dev->ifindex, state->dev->name,
		   NIP6(im->aca_addr),
		   im->aca_users);
	return 0;
}

static struct seq_operations ac6_seq_ops = {
	.start	=	ac6_seq_start,
	.next	=	ac6_seq_next,
	.stop	=	ac6_seq_stop,
	.show	=	ac6_seq_show,
};

static int ac6_seq_open(struct inode *inode, struct file *file)
{
	struct seq_file *seq;
	int rc = -ENOMEM;
	struct ac6_iter_state *s = kzalloc(sizeof(*s), GFP_KERNEL);

	if (!s)
		goto out;

	rc = seq_open(file, &ac6_seq_ops);
	if (rc)
		goto out_kfree;

	seq = file->private_data;
	seq->private = s;
out:
	return rc;
out_kfree:
	kfree(s);
	goto out;
}

static const struct file_operations ac6_seq_fops = {
	.owner		=	THIS_MODULE,
	.open		=	ac6_seq_open,
	.read		=	seq_read,
	.llseek		=	seq_lseek,
	.release	=	seq_release_private,
};

int __init ac6_proc_init(void)
{
	if (!proc_net_fops_create("anycast6", S_IRUGO, &ac6_seq_fops))
		return -ENOMEM;

	return 0;
}

void ac6_proc_exit(void)
{
	proc_net_remove("anycast6");
}
#endif
Commit	Line	Data
1da177e4 LT	1	/*
1da177e4 LT	2	* Anycast support for IPv6
1ab1457c	3	* Linux INET6 implementation
1da177e4 LT	4	*
	5	* Authors:
	6	* David L Stevens (dlstevens@us.ibm.com)
	7	*
	8	* based heavily on net/ipv6/mcast.c
	9	*
	10	* This program is free software; you can redistribute it and/or
	11	* modify it under the terms of the GNU General Public License
	12	* as published by the Free Software Foundation; either version
	13	* 2 of the License, or (at your option) any later version.
	14	*/
	15
4fc268d2	16	#include <linux/capability.h>
1da177e4 LT	17	#include <linux/module.h>
	18	#include <linux/errno.h>
	19	#include <linux/types.h>
	20	#include <linux/random.h>
	21	#include <linux/string.h>
	22	#include <linux/socket.h>
	23	#include <linux/sockios.h>
1da177e4 LT	24	#include <linux/net.h>
	25	#include <linux/in6.h>
	26	#include <linux/netdevice.h>
	27	#include <linux/if_arp.h>
	28	#include <linux/route.h>
	29	#include <linux/init.h>
	30	#include <linux/proc_fs.h>
	31	#include <linux/seq_file.h>
	32
	33	#include <net/sock.h>
	34	#include <net/snmp.h>
	35
	36	#include <net/ipv6.h>
	37	#include <net/protocol.h>
	38	#include <net/if_inet6.h>
	39	#include <net/ndisc.h>
	40	#include <net/addrconf.h>
	41	#include <net/ip6_route.h>
	42
	43	#include <net/checksum.h>
	44
	45	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr);
	46
	47	/* Big ac list lock for all the sockets */
	48	static DEFINE_RWLOCK(ipv6_sk_ac_lock);
	49
	50	static int
	51	ip6_onlink(struct in6_addr addr, struct net_device dev)
	52	{
	53	struct inet6_dev *idev;
	54	struct inet6_ifaddr *ifa;
	55	int onlink;
	56
	57	onlink = 0;
8814c4b5	58	rcu_read_lock();
1da177e4 LT	59	idev = __in6_dev_get(dev);
	60	if (idev) {
	61	read_lock_bh(&idev->lock);
	62	for (ifa=idev->addr_list; ifa; ifa=ifa->if_next) {
	63	onlink = ipv6_prefix_equal(addr, &ifa->addr,
	64	ifa->prefix_len);
	65	if (onlink)
	66	break;
	67	}
	68	read_unlock_bh(&idev->lock);
	69	}
8814c4b5	70	rcu_read_unlock();
1da177e4 LT	71	return onlink;
	72	}
	73
	74	/*
	75	* socket join an anycast group
	76	*/
	77
	78	int ipv6_sock_ac_join(struct sock sk, int ifindex, struct in6_addr addr)
	79	{
	80	struct ipv6_pinfo *np = inet6_sk(sk);
	81	struct net_device *dev = NULL;
	82	struct inet6_dev *idev;
	83	struct ipv6_ac_socklist *pac;
	84	int ishost = !ipv6_devconf.forwarding;
	85	int err = 0;
	86
	87	if (!capable(CAP_NET_ADMIN))
	88	return -EPERM;
	89	if (ipv6_addr_is_multicast(addr))
	90	return -EINVAL;
	91	if (ipv6_chk_addr(addr, NULL, 0))
	92	return -EINVAL;
	93
	94	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
	95	if (pac == NULL)
	96	return -ENOMEM;
	97	pac->acl_next = NULL;
	98	ipv6_addr_copy(&pac->acl_addr, addr);
	99
	100	if (ifindex == 0) {
	101	struct rt6_info *rt;
	102
	103	rt = rt6_lookup(addr, NULL, 0, 0);
	104	if (rt) {
	105	dev = rt->rt6i_dev;
	106	dev_hold(dev);
	107	dst_release(&rt->u.dst);
	108	} else if (ishost) {
	109	err = -EADDRNOTAVAIL;
	110	goto out_free_pac;
	111	} else {
	112	/* router, no matching interface: just pick one */
	113
	114	dev = dev_get_by_flags(IFF_UP, IFF_UP\|IFF_LOOPBACK);
	115	}
	116	} else
	117	dev = dev_get_by_index(ifindex);
	118
	119	if (dev == NULL) {
	120	err = -ENODEV;
	121	goto out_free_pac;
	122	}
	123
	124	idev = in6_dev_get(dev);
	125	if (!idev) {
	126	if (ifindex)
	127	err = -ENODEV;
	128	else
	129	err = -EADDRNOTAVAIL;
	130	goto out_dev_put;
	131	}
	132	/* reset ishost, now that we have a specific device */
	133	ishost = !idev->cnf.forwarding;
	134	in6_dev_put(idev);
135
136	pac->acl_ifindex = dev->ifindex;
137
138	/* XXX
139	* For hosts, allow link-local or matching prefix anycasts.
140	* This obviates the need for propagating anycast routes while
141	* still allowing some non-router anycast participation.
142	*/
143	if (!ip6_onlink(addr, dev)) {
144	if (ishost)
145	err = -EADDRNOTAVAIL;
146	if (err)
147	goto out_dev_put;
148	}
149
150	err = ipv6_dev_ac_inc(dev, addr);
151	if (err)
152	goto out_dev_put;
153
154	write_lock_bh(&ipv6_sk_ac_lock);
155	pac->acl_next = np->ipv6_ac_list;
156	np->ipv6_ac_list = pac;
157	write_unlock_bh(&ipv6_sk_ac_lock);
158
159	dev_put(dev);
160
161	return 0;
162
163	out_dev_put:
164	dev_put(dev);
165	out_free_pac:
166	sock_kfree_s(sk, pac, sizeof(*pac));
167	return err;
168	}
169
170	/*
171	* socket leave an anycast group
172	*/
173	int ipv6_sock_ac_drop(struct sock sk, int ifindex, struct in6_addr addr)
174	{
175	struct ipv6_pinfo *np = inet6_sk(sk);
176	struct net_device *dev;
177	struct ipv6_ac_socklist pac, prev_pac;
178
179	write_lock_bh(&ipv6_sk_ac_lock);
180	prev_pac = NULL;
181	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
182	if ((ifindex == 0 \|\| pac->acl_ifindex == ifindex) &&
183	ipv6_addr_equal(&pac->acl_addr, addr))
184	break;
185	prev_pac = pac;
186	}
187	if (!pac) {
188	write_unlock_bh(&ipv6_sk_ac_lock);
189	return -ENOENT;
190	}
191	if (prev_pac)
192	prev_pac->acl_next = pac->acl_next;
193	else
194	np->ipv6_ac_list = pac->acl_next;
195
196	write_unlock_bh(&ipv6_sk_ac_lock);
197
198	dev = dev_get_by_index(pac->acl_ifindex);
199	if (dev) {
200	ipv6_dev_ac_dec(dev, &pac->acl_addr);
201	dev_put(dev);
202	}
203	sock_kfree_s(sk, pac, sizeof(*pac));
204	return 0;
205	}
206
207	void ipv6_sock_ac_close(struct sock *sk)
208	{
209	struct ipv6_pinfo *np = inet6_sk(sk);
210	struct net_device *dev = NULL;
211	struct ipv6_ac_socklist *pac;
212	int prev_index;
213
214	write_lock_bh(&ipv6_sk_ac_lock);
215	pac = np->ipv6_ac_list;
216	np->ipv6_ac_list = NULL;
217	write_unlock_bh(&ipv6_sk_ac_lock);
218
219	prev_index = 0;
220	while (pac) {
221	struct ipv6_ac_socklist *next = pac->acl_next;
222
223	if (pac->acl_ifindex != prev_index) {
224	if (dev)
225	dev_put(dev);
226	dev = dev_get_by_index(pac->acl_ifindex);
227	prev_index = pac->acl_ifindex;
228	}
229	if (dev)
230	ipv6_dev_ac_dec(dev, &pac->acl_addr);
231	sock_kfree_s(sk, pac, sizeof(*pac));
232	pac = next;
233	}
234	if (dev)
235	dev_put(dev);
236	}
237
238	#if 0
239	/* The function is not used, which is funny. Apparently, author
240	* supposed to use it to filter out datagrams inside udp/raw but forgot.
241	*
242	* It is OK, anycasts are not special comparing to delivery to unicasts.
243	*/
244
245	int inet6_ac_check(struct sock sk, struct in6_addr addr, int ifindex)
246	{
247	struct ipv6_ac_socklist *pac;
248	struct ipv6_pinfo *np = inet6_sk(sk);
249	int found;
250
251	found = 0;
252	read_lock(&ipv6_sk_ac_lock);
253	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
254	if (ifindex && pac->acl_ifindex != ifindex)
255	continue;
256	found = ipv6_addr_equal(&pac->acl_addr, addr);
257	if (found)
258	break;
259	}
260	read_unlock(&ipv6_sk_ac_lock);
261
262	return found;
263	}
264
265	#endif
266
267	static void aca_put(struct ifacaddr6 *ac)
268	{
269	if (atomic_dec_and_test(&ac->aca_refcnt)) {
270	in6_dev_put(ac->aca_idev);
271	dst_release(&ac->aca_rt->u.dst);
272	kfree(ac);
273	}
274	}
275
276	/*
277	* device anycast group inc (add if not found)
278	*/
279	int ipv6_dev_ac_inc(struct net_device dev, struct in6_addr addr)
280	{
281	struct ifacaddr6 *aca;
282	struct inet6_dev *idev;
283	struct rt6_info *rt;
284	int err;
285
286	idev = in6_dev_get(dev);
287
288	if (idev == NULL)
289	return -EINVAL;
290
291	write_lock_bh(&idev->lock);
292	if (idev->dead) {
293	err = -ENODEV;
294	goto out;
295	}
296
297	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
298	if (ipv6_addr_equal(&aca->aca_addr, addr)) {
299	aca->aca_users++;
300	err = 0;
301	goto out;
302	}
303	}
304
305	/*
306	* not found: create a new one.
307	*/
308
0c600eda	309	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);
1da177e4 LT	310
	311	if (aca == NULL) {
	312	err = -ENOMEM;
	313	goto out;
	314	}
	315
	316	rt = addrconf_dst_alloc(idev, addr, 1);
	317	if (IS_ERR(rt)) {
	318	kfree(aca);
	319	err = PTR_ERR(rt);
	320	goto out;
	321	}
	322
1da177e4 LT	323	ipv6_addr_copy(&aca->aca_addr, addr);
	324	aca->aca_idev = idev;
	325	aca->aca_rt = rt;
	326	aca->aca_users = 1;
	327	/* aca_tstamp should be updated upon changes */
	328	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	329	atomic_set(&aca->aca_refcnt, 2);
	330	spin_lock_init(&aca->aca_lock);
	331
	332	aca->aca_next = idev->ac_list;
	333	idev->ac_list = aca;
	334	write_unlock_bh(&idev->lock);
	335
	336	dst_hold(&rt->u.dst);
40e22e8f	337	if (ip6_ins_rt(rt))
1da177e4 LT	338	dst_release(&rt->u.dst);
	339
	340	addrconf_join_solict(dev, &aca->aca_addr);
	341
	342	aca_put(aca);
	343	return 0;
	344	out:
	345	write_unlock_bh(&idev->lock);
	346	in6_dev_put(idev);
	347	return err;
	348	}
	349
	350	/*
	351	* device anycast group decrement
	352	*/
	353	int __ipv6_dev_ac_dec(struct inet6_dev idev, struct in6_addr addr)
	354	{
	355	struct ifacaddr6 aca, prev_aca;
	356
	357	write_lock_bh(&idev->lock);
	358	prev_aca = NULL;
	359	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	360	if (ipv6_addr_equal(&aca->aca_addr, addr))
	361	break;
	362	prev_aca = aca;
	363	}
	364	if (!aca) {
	365	write_unlock_bh(&idev->lock);
	366	return -ENOENT;
	367	}
	368	if (--aca->aca_users > 0) {
	369	write_unlock_bh(&idev->lock);
	370	return 0;
	371	}
	372	if (prev_aca)
	373	prev_aca->aca_next = aca->aca_next;
	374	else
	375	idev->ac_list = aca->aca_next;
	376	write_unlock_bh(&idev->lock);
	377	addrconf_leave_solict(idev, &aca->aca_addr);
	378
	379	dst_hold(&aca->aca_rt->u.dst);
72d3b2c9	380	if (ip6_del_rt(aca->aca_rt))
1da177e4 LT	381	dst_free(&aca->aca_rt->u.dst);
	382	else
	383	dst_release(&aca->aca_rt->u.dst);
	384
	385	aca_put(aca);
	386	return 0;
	387	}
	388
	389	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr)
	390	{
	391	int ret;
	392	struct inet6_dev *idev = in6_dev_get(dev);
	393	if (idev == NULL)
	394	return -ENODEV;
	395	ret = __ipv6_dev_ac_dec(idev, addr);
	396	in6_dev_put(idev);
	397	return ret;
	398	}
1ab1457c	399
1da177e4 LT	400	/*
	401	* check if the interface has this anycast address
	402	*/
	403	static int ipv6_chk_acast_dev(struct net_device dev, struct in6_addr addr)
	404	{
	405	struct inet6_dev *idev;
	406	struct ifacaddr6 *aca;
	407
	408	idev = in6_dev_get(dev);
	409	if (idev) {
	410	read_lock_bh(&idev->lock);
	411	for (aca = idev->ac_list; aca; aca = aca->aca_next)
	412	if (ipv6_addr_equal(&aca->aca_addr, addr))
	413	break;
	414	read_unlock_bh(&idev->lock);
	415	in6_dev_put(idev);
	416	return aca != 0;
	417	}
	418	return 0;
	419	}
	420
	421	/*
	422	* check if given interface (or any, if dev==0) has this anycast address
	423	*/
	424	int ipv6_chk_acast_addr(struct net_device dev, struct in6_addr addr)
	425	{
	426	if (dev)
	427	return ipv6_chk_acast_dev(dev, addr);
	428	read_lock(&dev_base_lock);
	429	for (dev=dev_base; dev; dev=dev->next)
	430	if (ipv6_chk_acast_dev(dev, addr))
	431	break;
	432	read_unlock(&dev_base_lock);
	433	return dev != 0;
	434	}
	435
	436
	437	#ifdef CONFIG_PROC_FS
	438	struct ac6_iter_state {
	439	struct net_device *dev;
	440	struct inet6_dev *idev;
	441	};
	442
	443	#define ac6_seq_private(seq) ((struct ac6_iter_state *)(seq)->private)
	444
	445	static inline struct ifacaddr6 ac6_get_first(struct seq_file seq)
	446	{
	447	struct ifacaddr6 *im = NULL;
	448	struct ac6_iter_state *state = ac6_seq_private(seq);
	449
	450	for (state->dev = dev_base, state->idev = NULL;
	451	state->dev;
	452	state->dev = state->dev->next) {
	453	struct inet6_dev *idev;
	454	idev = in6_dev_get(state->dev);
	455	if (!idev)
	456	continue;
	457	read_lock_bh(&idev->lock);
	458	im = idev->ac_list;
	459	if (im) {
	460	state->idev = idev;
	461	break;
	462	}
	463	read_unlock_bh(&idev->lock);
464	}
465	return im;
466	}
467
468	static struct ifacaddr6 ac6_get_next(struct seq_file seq, struct ifacaddr6 *im)
469	{
470	struct ac6_iter_state *state = ac6_seq_private(seq);
471
472	im = im->aca_next;
473	while (!im) {
474	if (likely(state->idev != NULL)) {
475	read_unlock_bh(&state->idev->lock);
476	in6_dev_put(state->idev);
477	}
478	state->dev = state->dev->next;
479	if (!state->dev) {
480	state->idev = NULL;
481	break;
482	}
483	state->idev = in6_dev_get(state->dev);
484	if (!state->idev)
485	continue;
486	read_lock_bh(&state->idev->lock);
487	im = state->idev->ac_list;
488	}
489	return im;
490	}
491
492	static struct ifacaddr6 ac6_get_idx(struct seq_file seq, loff_t pos)
493	{
494	struct ifacaddr6 *im = ac6_get_first(seq);
495	if (im)
496	while (pos && (im = ac6_get_next(seq, im)) != NULL)
497	--pos;
498	return pos ? NULL : im;
499	}
500
501	static void ac6_seq_start(struct seq_file seq, loff_t *pos)
502	{
503	read_lock(&dev_base_lock);
504	return ac6_get_idx(seq, *pos);
505	}
506
507	static void ac6_seq_next(struct seq_file seq, void v, loff_t pos)
508	{
509	struct ifacaddr6 *im;
510	im = ac6_get_next(seq, v);
511	++*pos;
512	return im;
513	}
514
515	static void ac6_seq_stop(struct seq_file seq, void v)
516	{
517	struct ac6_iter_state *state = ac6_seq_private(seq);
518	if (likely(state->idev != NULL)) {
519	read_unlock_bh(&state->idev->lock);
520	in6_dev_put(state->idev);
521	}
522	read_unlock(&dev_base_lock);
523	}
524
525	static int ac6_seq_show(struct seq_file seq, void v)
526	{
527	struct ifacaddr6 im = (struct ifacaddr6 )v;
528	struct ac6_iter_state *state = ac6_seq_private(seq);
529
530	seq_printf(seq,
9343e79a	531	"%-4d %-15s " NIP6_SEQFMT " %5d\n",
1da177e4 LT	532	state->dev->ifindex, state->dev->name,
	533	NIP6(im->aca_addr),
	534	im->aca_users);
	535	return 0;
	536	}
	537
	538	static struct seq_operations ac6_seq_ops = {
	539	.start = ac6_seq_start,
	540	.next = ac6_seq_next,
	541	.stop = ac6_seq_stop,
	542	.show = ac6_seq_show,
	543	};
	544
	545	static int ac6_seq_open(struct inode inode, struct file file)
	546	{
	547	struct seq_file *seq;
	548	int rc = -ENOMEM;
0c600eda	549	struct ac6_iter_state s = kzalloc(sizeof(s), GFP_KERNEL);
1da177e4 LT	550
	551	if (!s)
	552	goto out;
	553
	554	rc = seq_open(file, &ac6_seq_ops);
	555	if (rc)
	556	goto out_kfree;
	557
	558	seq = file->private_data;
	559	seq->private = s;
1da177e4 LT	560	out:
	561	return rc;
	562	out_kfree:
	563	kfree(s);
	564	goto out;
	565	}
	566
9a32144e	567	static const struct file_operations ac6_seq_fops = {
1da177e4 LT	568	.owner = THIS_MODULE,
	569	.open = ac6_seq_open,
	570	.read = seq_read,
	571	.llseek = seq_lseek,
	572	.release = seq_release_private,
	573	};
	574
	575	int __init ac6_proc_init(void)
	576	{
	577	if (!proc_net_fops_create("anycast6", S_IRUGO, &ac6_seq_fops))
	578	return -ENOMEM;
	579
	580	return 0;
	581	}
	582
	583	void ac6_proc_exit(void)
	584	{
	585	proc_net_remove("anycast6");
	586	}
	587	#endif
	588